diff --git a/examples/onnx/159008.jpg b/examples/onnx/159008.jpg
new file mode 100644
index 0000000..2b90845
--- /dev/null
+++ b/examples/onnx/159008.jpg
Binary files differ
diff --git a/examples/onnx/densenet121.py b/examples/onnx/densenet121.py
index 14cfa5f..c8a8f95 100644
--- a/examples/onnx/densenet121.py
+++ b/examples/onnx/densenet121.py
@@ -22,10 +22,9 @@
 
 from singa import device
 from singa import tensor
-from singa import autograd
 from singa import sonnx
 import onnx
-from utils import download_model, update_batch_size, check_exist_or_download
+from utils import download_model, check_exist_or_download
 
 import logging
 logging.basicConfig(level=logging.INFO, format='%(asctime)-15s %(message)s')
@@ -44,7 +43,7 @@
     return img
 
 
-def get_image_labe():
+def get_image_label():
     # download label
     label_url = 'https://s3.amazonaws.com/onnx-model-zoo/synset.txt'
     with open(check_exist_or_download(label_url), 'r') as f:
@@ -56,56 +55,52 @@
     return img, labels
 
 
-class Infer:
+class MyModel(sonnx.SONNXModel):
 
-    def __init__(self, sg_ir):
-        self.sg_ir = sg_ir
-        for idx, tens in sg_ir.tensor_map.items():
-            # allow the tensors to be updated
-            tens.requires_grad = True
-            tens.stores_grad = True
-            sg_ir.tensor_map[idx] = tens
+    def __init__(self, onnx_model):
+        super(MyModel, self).__init__(onnx_model)
 
-    def forward(self, x):
-        return sg_ir.run([x])[0]
+    def forward(self, *x):
+        y = super(MyModel, self).forward(*x)
+        return y[0]
+
+    def train_one_batch(self, x, y):
+        pass
 
 
 if __name__ == "__main__":
+
+    download_dir = '/tmp'
     url = 'https://s3.amazonaws.com/download.onnx/models/opset_9/densenet121.tar.gz'
-    download_dir = '/tmp/'
     model_path = os.path.join(download_dir, 'densenet121', 'model.onnx')
 
     logging.info("onnx load model...")
     download_model(url)
     onnx_model = onnx.load(model_path)
 
-    # set batch size
-    onnx_model = update_batch_size(onnx_model, 1)
+    # inference demo
+    logging.info("preprocessing...")
+    img, labels = get_image_label()
+    img = preprocess(img)
+    # sg_ir = sonnx.prepare(onnx_model) # run without graph
+    # y = sg_ir.run([img])
 
-    # prepare the model
-    logging.info("prepare model...")
+    logging.info("model compling...")
     dev = device.create_cuda_gpu()
-    sg_ir = sonnx.prepare(onnx_model, device=dev)
-    autograd.training = False
-    model = Infer(sg_ir)
+    x = tensor.Tensor(device=dev, data=img)
+    model = MyModel(onnx_model)
+    model.compile([x], is_train=False, use_graph=True, sequential=True)
 
     # verifty the test
     # from utils import load_dataset
-    # inputs, ref_outputs = load_dataset(
-    #     os.path.join('/tmp', 'densenet121', 'test_data_set_0'))
+    # inputs, ref_outputs = load_dataset(os.path.join('/tmp', 'densenet121', 'test_data_set_0'))
     # x_batch = tensor.Tensor(device=dev, data=inputs[0])
-    # outputs = model.forward(x_batch)
+    # outputs = sg_ir.run([x_batch])
     # for ref_o, o in zip(ref_outputs, outputs):
     #     np.testing.assert_almost_equal(ref_o, tensor.to_numpy(o), 4)
 
-    # inference
-    logging.info("preprocessing...")
-    img, labels = get_image_labe()
-    img = preprocess(img)
-
     logging.info("model running...")
-    x_batch = tensor.Tensor(device=dev, data=img)
-    y = model.forward(x_batch)
+    y = model.forward(x)
 
     logging.info("postprocessing...")
     y = tensor.softmax(y)
@@ -113,4 +108,4 @@
     scores = np.squeeze(scores)
     a = np.argsort(scores)[::-1]
     for i in a[0:5]:
-        logging.info('class=%s ; probability=%f' % (labels[i], scores[i]))
+        logging.info('class=%s ; probability=%f' % (labels[i], scores[i]))
\ No newline at end of file
diff --git a/examples/onnx/shufflenetv1.py b/examples/onnx/shufflenetv1.py
index 35fa777..139d6d3 100644
--- a/examples/onnx/shufflenetv1.py
+++ b/examples/onnx/shufflenetv1.py
@@ -23,12 +23,9 @@
 
 from singa import device
 from singa import tensor
-from singa import autograd
 from singa import sonnx
 import onnx
-from utils import download_model
-from utils import update_batch_size
-from utils import check_exist_or_download
+from utils import download_model, check_exist_or_download
 
 logging.basicConfig(level=logging.INFO, format='%(asctime)s %(message)s')
 
@@ -56,51 +53,53 @@
     return img, labels
 
 
-class Infer:
+class MyModel(sonnx.SONNXModel):
 
-    def __init__(self, sg_ir):
-        self.sg_ir = sg_ir
-        for idx, tens in sg_ir.tensor_map.items():
-            tens.require_grad = True
-            tens.store_grad = True
-            sg_ir.tensor_map[idx] = tens
+    def __init__(self, onnx_model):
+        super(MyModel, self).__init__(onnx_model)
 
-    def forward(self, x):
-        return sg_ir.run([x])[0]
+    def forward(self, *x):
+        y = super(MyModel, self).forward(*x)
+        return y[0]
+
+    def train_one_batch(self, x, y):
+        pass
 
 
 if __name__ == '__main__':
+    download_dir = '/tmp'
     url = 'https://github.com/onnx/models/raw/master/vision/classification/shufflenet/model/shufflenet-9.tar.gz'
-    download_dir = "/tmp/"
     model_path = os.path.join(download_dir, 'shufflenet', 'model.onnx')
-    logging.info("onnx load model....")
+
+    logging.info("onnx load model...")
     download_model(url)
     onnx_model = onnx.load(model_path)
-    # setting batch size
-    onnx_model = update_batch_size(onnx_model, 1)
-    # preparing the model
-    logging.info("preparing model...")
-    dev = device.create_cuda_gpu()
-    sg_ir = sonnx.prepare(onnx_model, device=dev)
-    autograd.training = False
-    model = Infer(sg_ir)
-	
-    # verifying the test dataset
-    #from utils import load_dataset
-    #inputs,ref_outputs = load_dataset(os.path.join('/tmp','shufflenet','test_data_set_0'))
-    #x_batch = tensor.Tensor(device = dev,data=inputs[0])
-    #outputs = model.forward(x_batch)
-    # for ref_o,o in zip(ref_outputs,outputs):
-    #    np.testing.assert_almost_equal(ref_o,tensor.to_numpy(o),4)
-    
-    # inference
+
+    # inference demo
     logging.info("preprocessing...")
     img, labels = get_image_label()
     img = preprocess(img)
-    x_batch = tensor.Tensor(device=dev, data=img)
-    logging.info("model running....")
-    y = model.forward(x_batch)
-    logging.info("postprocessing....")
+    # sg_ir = sonnx.prepare(onnx_model) # run without graph
+    # y = sg_ir.run([img])
+
+    logging.info("model compling...")
+    dev = device.create_cuda_gpu()
+    x = tensor.Tensor(device=dev, data=img)
+    model = MyModel(onnx_model)
+    model.compile([x], is_train=False, use_graph=True, sequential=True)
+
+    # verifty the test
+    # from utils import load_dataset
+    # inputs, ref_outputs = load_dataset(os.path.join('/tmp', 'shufflenet', 'test_data_set_0'))
+    # x_batch = tensor.Tensor(device=dev, data=inputs[0])
+    # outputs = sg_ir.run([x_batch])
+    # for ref_o, o in zip(ref_outputs, outputs):
+    #     np.testing.assert_almost_equal(ref_o, tensor.to_numpy(o), 4)
+
+    logging.info("model running...")
+    y = model.forward(x)
+
+    logging.info("postprocessing...")
     y = tensor.softmax(y)
     scores = tensor.to_numpy(y)
     scores = np.squeeze(scores)
diff --git a/examples/onnx/shufflenetv2.py b/examples/onnx/shufflenetv2.py
index 74dd794..60f84a4 100644
--- a/examples/onnx/shufflenetv2.py
+++ b/examples/onnx/shufflenetv2.py
@@ -43,7 +43,7 @@
     return img
 
 
-def get_image_labe():
+def get_image_label():
     # download label
     label_url = 'https://s3.amazonaws.com/onnx-model-zoo/synset.txt'
     with open(check_exist_or_download(label_url), 'r') as f:
@@ -81,7 +81,7 @@
 
     # inference
     logging.info("preprocessing...")
-    img, labels = get_image_labe()
+    img, labels = get_image_label()
     img = preprocess(img)
     # sg_ir = sonnx.prepare(onnx_model) # run without graph
     # y = sg_ir.run([img])
diff --git a/examples/onnx/squeezenet.py b/examples/onnx/squeezenet.py
index 8a6ecf5..861550c 100644
--- a/examples/onnx/squeezenet.py
+++ b/examples/onnx/squeezenet.py
@@ -22,10 +22,9 @@
 
 from singa import device
 from singa import tensor
-from singa import autograd
 from singa import sonnx
 import onnx
-from utils import download_model, update_batch_size, check_exist_or_download
+from utils import download_model, check_exist_or_download
 
 import logging
 logging.basicConfig(level=logging.INFO, format='%(asctime)-15s %(message)s')
@@ -56,24 +55,22 @@
     return img, labels
 
 
-class Infer:
+class MyModel(sonnx.SONNXModel):
 
-    def __init__(self, sg_ir):
-        self.sg_ir = sg_ir
-        for idx, tens in sg_ir.tensor_map.items():
-            # allow the tensors to be updated
-            tens.requires_grad = True
-            tens.stores_grad = True
-            sg_ir.tensor_map[idx] = tens
+    def __init__(self, onnx_model):
+        super(MyModel, self).__init__(onnx_model)
 
-    def forward(self, x):
-        return sg_ir.run([x])[0]
+    def forward(self, *x):
+        y = super(MyModel, self).forward(*x)
+        return y[0]
+
+    def train_one_batch(self, x, y):
+        pass
 
 
-if __name__ == "__main__":
-
+if __name__ == '__main__':
+    download_dir = '/tmp'
     url = 'https://github.com/onnx/models/raw/master/vision/classification/squeezenet/model/squeezenet1.1-7.tar.gz'
-    download_dir = '/tmp/'
     model_path = os.path.join(download_dir, 'squeezenet1.1',
                               'squeezenet1.1.onnx')
 
@@ -81,33 +78,29 @@
     download_model(url)
     onnx_model = onnx.load(model_path)
 
-    # set batch size
-    onnx_model = update_batch_size(onnx_model, 1)
-
-    # prepare the model
-    logging.info("prepare model...")
-    dev = device.create_cuda_gpu()
-    sg_ir = sonnx.prepare(onnx_model, device=dev)
-    autograd.training = False
-    model = Infer(sg_ir)
-
-    # verify the test
-    # from utils import load_dataset
-    # inputs, ref_outputs = load_dataset(
-    #     os.path.join('/tmp', 'squeezenet1.1', 'test_data_set_0'))
-    # x_batch = tensor.Tensor(device=dev, data=inputs[0])
-    # outputs = model.forward(x_batch)
-    # for ref_o, o in zip(ref_outputs, outputs):
-    #     np.testing.assert_almost_equal(ref_o, tensor.to_numpy(o), 4)
-
-    # inference
+    # inference demo
     logging.info("preprocessing...")
     img, labels = get_image_label()
     img = preprocess(img)
+    # sg_ir = sonnx.prepare(onnx_model) # run without graph
+    # y = sg_ir.run([img])
+
+    logging.info("model compling...")
+    dev = device.create_cuda_gpu()
+    x = tensor.Tensor(device=dev, data=img)
+    model = MyModel(onnx_model)
+    model.compile([x], is_train=False, use_graph=True, sequential=True)
+
+    # verifty the test
+    # from utils import load_dataset
+    # inputs, ref_outputs = load_dataset(os.path.join('/tmp', 'squeezenet1.1', 'test_data_set_0'))
+    # x_batch = tensor.Tensor(device=dev, data=inputs[0])
+    # outputs = sg_ir.run([x_batch])
+    # for ref_o, o in zip(ref_outputs, outputs):
+    #     np.testing.assert_almost_equal(ref_o, tensor.to_numpy(o), 4)
 
     logging.info("model running...")
-    x_batch = tensor.Tensor(device=dev, data=img)
-    y = model.forward(x_batch)
+    y = model.forward(x)
 
     logging.info("postprocessing...")
     y = tensor.softmax(y)
diff --git a/examples/onnx/superresolution.py b/examples/onnx/superresolution.py
index 0e752ba..caaae19 100644
--- a/examples/onnx/superresolution.py
+++ b/examples/onnx/superresolution.py
@@ -19,7 +19,6 @@
 import os
 import numpy as np
 from PIL import Image
-# pip install python-resize-image
 from resizeimage import resizeimage
 
 from singa import device
diff --git a/examples/onnx/vgg19.py b/examples/onnx/vgg19.py
index 49606cb..a2c3ea7 100644
--- a/examples/onnx/vgg19.py
+++ b/examples/onnx/vgg19.py
@@ -22,10 +22,9 @@
 
 from singa import device
 from singa import tensor
-from singa import autograd
 from singa import sonnx
 import onnx
-from utils import download_model, update_batch_size, check_exist_or_download
+from utils import download_model, check_exist_or_download
 
 import logging
 logging.basicConfig(level=logging.INFO, format='%(asctime)-15s %(message)s')
@@ -44,7 +43,7 @@
     return img
 
 
-def get_image_label():
+def get_image_labe():
     # download label
     label_url = 'https://s3.amazonaws.com/onnx-model-zoo/synset.txt'
     with open(check_exist_or_download(label_url), 'r') as f:
@@ -56,22 +55,21 @@
     return img, labels
 
 
-class Infer:
+class MyModel(sonnx.SONNXModel):
 
-    def __init__(self, sg_ir):
-        self.sg_ir = sg_ir
-        for idx, tens in sg_ir.tensor_map.items():
-            # allow the tensors to be updated
-            tens.requires_grad = True
-            tens.stores_grad = True
-            sg_ir.tensor_map[idx] = tens
+    def __init__(self, onnx_model):
+        super(MyModel, self).__init__(onnx_model)
 
-    def forward(self, x):
-        return sg_ir.run([x])[0]
+    def forward(self, *x):
+        y = super(MyModel, self).forward(*x)
+        return y[0]
+
+    def train_one_batch(self, x, y):
+        pass
 
 
 if __name__ == "__main__":
-    url = 'https://github.com/onnx/models/raw/master/vision/classification/vgg/model/vgg19-7.tar.gz'
+    url = 'https://s3.amazonaws.com/onnx-model-zoo/vgg/vgg19/vgg19.tar.gz'
     download_dir = '/tmp/'
     model_path = os.path.join(download_dir, 'vgg19', 'vgg19.onnx')
 
@@ -79,33 +77,30 @@
     download_model(url)
     onnx_model = onnx.load(model_path)
 
-    # set batch size
-    onnx_model = update_batch_size(onnx_model, 1)
+    # inference
+    logging.info("preprocessing...")
+    img, labels = get_image_labe()
+    img = preprocess(img)
+    # sg_ir = sonnx.prepare(onnx_model) # run without graph
+    # y = sg_ir.run([img])
 
-    # prepare the model
-    logging.info("prepare model...")
-    # dev = device.get_default_device()
+    logging.info("model compling...")
     dev = device.create_cuda_gpu()
-    sg_ir = sonnx.prepare(onnx_model, device=dev)
-    autograd.training = False
-    model = Infer(sg_ir)
+    x = tensor.PlaceHolder(img.shape, device=dev)
+    model = MyModel(onnx_model)
+    model.compile([x], is_train=False, use_graph=True, sequential=True)
 
-    # verify the test
+    # verifty the test
     # from utils import load_dataset
     # inputs, ref_outputs = load_dataset(os.path.join('/tmp', 'vgg19', 'test_data_set_0'))
     # x_batch = tensor.Tensor(device=dev, data=inputs[0])
-    # outputs = model.forward(x_batch)
+    # outputs = sg_ir.run([x_batch])
     # for ref_o, o in zip(ref_outputs, outputs):
     #     np.testing.assert_almost_equal(ref_o, tensor.to_numpy(o), 4)
 
-    # inference
-    logging.info("preprocessing...")
-    img, labels = get_image_label()
-    img = preprocess(img)
-
     logging.info("model running...")
-    x_batch = tensor.Tensor(device=dev, data=img)
-    y = model.forward(x_batch)
+    x = tensor.Tensor(device=dev, data=img)
+    y = model.forward(x)
 
     logging.info("postprocessing...")
     y = tensor.softmax(y)
diff --git a/python/singa/utils.py b/python/singa/utils.py
index d102771..7dcf8fb 100644
--- a/python/singa/utils.py
+++ b/python/singa/utils.py
@@ -16,7 +16,6 @@
 # =============================================================================
 
 import sys
-import math
 import numpy as np
 import collections